Measures of response that are clinically meaningful and occur early are an unmet need in metastatic castration-resistant prostate cancer clinical research and practice. We explored, using individual patient data, week 13 circulating tumor cell (CTC) and prostate-specific antigen (PSA) response end points in five prospective randomized phase III trials that enrolled a total of 6,081 patients-COU-AA-301, AFFIRM, ELM-PC-5, ELM-PC-4, and COMET-1-ClinicalTrials.Gov identifiers: NCT00638690, NCT00974311, NCT01193257, NCT01193244, and NCT01605227, respectively.
MethodsEight response end points were explored. CTC nonzero at baseline and 0 at 13 weeks (CTC0); CTC conversion ($ 5 CTCs at baseline, # 4 at 13 weeks-the US Food and Drug Administration cleared response measure); a 30%, 50%, and 70% decrease in CTC count; and a 30%, 50%, and 70% decrease in PSA level. Patients missing week-13 values were considered nonresponders. The discriminatory strength of each end point with respect to overall survival in each trial was assessed using the weighted c-index.
ResultsOf the eight response end points, CTC0 and CTC conversion had the highest weighted c-indices, with smaller standard deviations. For CTC0, the mean (standard deviation) was 0.81 (0.04); for CTC conversion, 0.79 (0.03); for 30% decrease in CTC count, 0.72 (0.06); for 50% decrease in CTC count, 0.72 (0.06); for 70% decrease in CTC count, 0.73 (0.05); for 30% decrease in PSA level, 0.71 (0.03); for 50% decrease in PSA level, 0.72 (0.06); and for 70% decrease in PSA level, 0.74 (0.05). Seventyfive percent of eligible patients could be evaluated with the CTC0 end point, compared with 51% with the CTC conversion end point.
ConclusionThe CTC0 and CTC conversion end points had the highest discriminatory power for overall survival. Both are robust and meaningful response end points for early-phase metastatic castration-resistant prostate cancer clinical trials. CTC0 is applicable to a significantly higher percentage of patients than CTC conversion.